WO2003079457A1

WO2003079457A1 - Luminescence conversion and application to photovoltaic energy conversion

Info

Publication number: WO2003079457A1
Application number: PCT/AU2003/000333
Authority: WO
Inventors: Thorsten Trupke; Martin Andrew Green; Peter Wuerfel
Original assignee: Unisearch Limited
Priority date: 2002-03-19
Filing date: 2003-03-19
Publication date: 2003-09-25
Also published as: AUPS123302A0

Abstract

A system for solar energy conversion using the up or down conversion of sub-band-gap photons to increase the maximum efficiency of a single-junction conventional solar cell is discussed. Different embodiments show the up or down converter (15) between a solar cell (13) and a reflector (16). This structure may be doubled to include converters on both sides of the solar cell. Optional insulator (14) may be placed between the solar cell and the converter. The reflector reflects the radiation back to the solar cell and converter to allow more uptake of available radiation.

Description

Luminescence conversion and application to photovoltaic energy conversion Introduction

The present invention relates to the field of luminescence conversion particularly with reference to photovoltaics and in particular the invention provides improvements in luminescence converters and application of improved luminescence converters to photovoltaic devices such as solar cells and photodetectors Background of the Invention

The main problem in solar energy conversion is the fact that the incident solar spectrum consists of photons in a significant amount in a very broad spectral range up to roughly 5 eV Conventional photovoltaic devices are based on materials like semiconductors or dye molecules, which have a lower threshold energy for the absorptance The two major loss mechanisms leading to reduced energy conversion efficiencies in such conventional devices are 1) Transmission losses' Incident photons with energies smaller than the threshold energy for the absorptance are transmitted and cannot be used by the photovoltaic cell

2) Thermalisation losses The absorption of incident photons with energies larger than the threshold energy for the absorptance leads to the generation of only one electron-hole pair per absorbed photon, regardless of the photon energy The excess energy of an incident photon above the threshold energy is wasted during the thermalisation of the generated electron-hole pairs

One concept to reduce the transmission losses is the impurity photovoltaic effect

(IPV), i e the insertion of impurities with energies located in the band-gap, which was proposed by M Wolf in Proc IRE, 48, 1960, 1246 However, the introduction of impurities into a photovoltaic cell material has also a major disadvantage From detailed balance analysis it follows that the additional generation channel for electron hole pairs also represents an additional recombination channel A problem inherent in an IPV-system is that not only the additional electron hole pairs, which are gained by the insertion of impurities, but all electron hole pairs are affected by the additional recombination channel

Approaches to reduce the thermalisation losses include hot carrier solar cells and solar cells in which multiple electron-hole pair generation per incident photon is possible due to impact ionisation Significant improvements of the efficiency over the efficiencies of traditional photovoltaic cells are predicted theoretically for these systems However, the assumption of a largely reduced electron-phonon coupling, which must be made in the theoretical description of these systems are impossible to fulfil with present solar cell materials.

One thing all the above-mentioned approaches for improved solar cell systems have in common is that they try to adapt the solar cell material to the broad spectral range of the incident solar spectrum. However, the conditions on the material quality, which must be fulfilled in order to achieve improved energy conversion efficiencies with these approaches, are very unlikely to be fulfilled in the near future. Summary of the Invention

The present invention consists in a photovoltaic device comprising a photovoltaic cell, a luminescence converter and a reflector.

Instead of adapting the photovoltaic cell to the broad incident solar spectrum, embodiments of the present invention are proposed to change the incident spectrum before it is absorbed by the photovoltaic cell and converts photons, which cannot effectively be utilised by the photovoltaic cell to different energies. Two major advantages of the currently proposed embodiments of the invention are:

1.) Different constraints are imposed on the material quality for an efficient photovoltaic cell or for an efficient luminescence converter. Contrary to e.g. an IPV-solar cell the material of the luminescence converter can be optimised independently of the solar cell material itself.

2.) The new approaches can be applied, in principle, to any existing solar cell. In a first embodiment of the invention the transmission losses in a photovoltaic cell are reduced. The luminescence converter is used as an up-converter, which transforms sub-band-gap photons transmitted by the photovoltaic cell to higher energy- photons, which can subsequently be absorbed by the photovoltaic cell. The up- converter is located between the photovoltaic cell and a rear reflector and is optically coupled to the photovoltaic cell. The up-converter and the photovoltaic cell must be substantially electronically isolated from each other or at least only weakly coupled. This can be achieved by either fabricating the up-converter from material in which the electrons do not couple well to the photovoltaic cell material even when the up- converter and the photovoltaic cell are in direct mechanical contact, or by separating the up-converter and the photovoltaic cell, with a transparent insulating layer. The luminescence converter, the photovoltaic cell and if necessary also the insulating layer preferably all have the same or substantially similar refractive indices. The up-converter preferably consists of a system in which three bands are involved in optical transitions, where each band consists of allowed energy states of a single energy, or spread over a range of energies The states associated with each band may or may not be in electrical communication with one another The three bands are defined as "lower band", "intermediate band" and "upper band". One type of up- conversion processes, which is denoted ground state absorption/excited state absorption (GSA/ESA), involves the generation of an excited state by a two-step process involving a real, lower-lying, metastable state or states According to a theoretical model described below, the up-conversion efficiency involving these standard processes are very low for non-concentrated sunlight and only minor improvements are calculated even under the most idealising assumptions In a preferred form of the first embodiment the system includes an improved up-converter, which includes at least one excited energy state to which electrons can be excited from an initial lower state by absorption of an incident sub-band-gap energy photon and from which an electron will relax into a rest state which is lower than the excited state but higher than the initial state with consequent release of a small amount of energy and where the rest state is only weakly linked to the initial state from which the electron was originally excited The rest state can be associated with the upper band or an intermediate level band preferably or both of these bands can have an associated rest state

Gibart et al (P Gibart, F Auzel, J C Guillaume, K Zahraman, Jap J Appl Phys , 35, 1996, 4401) attempted to demonstrate experimentally the feasibility of the up-conversion of sub-band-gap light in photovoltaic devices by stacking a rare earth doped vitroceramic behind a substrate-free GaAs solar cell In their paper, Gibart et al conclude that a practical application of the up-conversion process in photovoltaics is not effective as the power conversion efficiency of their cell was only 2 5% Contrary to the conclusions of Gibart et al , it is shown that, theoretically, a photovoltaic system with a suitable up-converter with one intermediate level can have an efficiency of up to 63 17% for non-concentrated sunlight if the solid-angle, into which luminescent emission from the system takes place, is restricted to the solar solid angle Ω_s Without that restriction of the solid angle an efficiency of 47 6% is the limit on what can be achieved with an improved up-converter as described above

Preferred embodiments may incorporate a semiconductor or any other material with a band-type electronic density of states and with an intermediate level or intermediate band within the band gap where all transitions are radiatively efficient, to act as a broad band up-converter In a second embodiment of the invention the thermalisation losses in a photovoltaic cell can largely be reduced In this embodiment of the invention the luminescence converter is a down-converter which absorbs photons at an energy at or greater than twice the band gap of the photovoltaic cell and emits two or more photons at energies at or above the band-gap energy of the photovoltaic cell. These lower- energy photons can both be absorbed by the photovoltaic cell and contribute to the photocurrent.

In one preferred form of the second embodiment of the invention the down- converter is located in front of a photovoltaic cell. The converter absorbs photons with energies equal to or larger than twice the band-gap energy E_g of the photovoltaic device and can be almost transparent to photons with energies larger than E_g but smaller than 2E_g. The latter are not suitable for down-conversion and their absorption inside the down-converter would reduce the efficiency of the system.

Preferably, the luminescence converter and the photovoltaic cell are optically coupled to each other but electrically isolated from each other. If the materials of the down-converter and the photovoltaic cell are capable of coupling electronically to each other, then an interspaced transparent insulating must be located between the down- converter and the photovoltaic device. The luminescence converter, the photovoltaic cell and the insulating layer (if required) preferably all have the same or substantially similar refractive indices. The insulating layer is not required if the luminance converter and the photovoltaic cell do not couple electronically when directly mechanically attached to each other.

In another form of the second embodiment the down-converter is located on the rear surface of the photovoltaic cell. The theoretical description of the down- conversion system reveals that the highest limiting efficiency can be expected for this form of the second embodiment. An efficiency limit of 39.55% is calculated for a photovoltaic device in combination with a down-converter for a band-gap energy of the photovoltaic cell of E_g=l . l eV which makes this approach very attractive for silicon solar cells or silicon photo detectors. However, disadvantages of this form of the second embodiment are that a bifacial photovoltaic cell is required and that the photovoltaic cell must be transparent to those incident photons, which shall be down- converted, which rules out all semiconductor materials.

The ow?-conversion of incident photons into two or more photons with lower energies can be achieved by a material with a separation between the upper and lower band of states equal to at least 2.E_g, which contains states or bands of intermediate energies as for the up-converter case. Therefore a material similar in characteristics to that used for the previously discussed up-converter can also be used to convert high- energy photons into two lower energy photons. The absorption of high-energy photons by a lower band-to-band transition leads to the generation of electron-hole-pairs inside the luminescence converter. The radiative recombination of these electron-hole-pairs via the intermediate band is accompanied by the emission of two lower energy photons. Recently a growing interest in down-converting materials has arisen due to commercial applications for mercury-free fluorescent tubes and plasma display panels. R. T. Wegh, H. Donker, E. N. D. v. Loef, K. D. Oskam and A. Meijerink (Journal of Luminescence, 87-89, 2000, 1017) have demonstrated down conversion efficiencies close to 200% for conversion of ultraviolet light into visible light using different rare earth compounds. In a third embodiment of the invention an up-converter on the rear surface and a down-converter on the front surface of a solar cell are combined to reduce both, the transmission losses and the thermalisation losses.

The improvements of photovoltaic cell efficiencies due to the above mentioned approaches involving luminescence converters are achieved by increased generation rates for electron-hole pairs. Another interesting application for luminescence converters is the use of the proposed processes in photo-detectors. The sensitivity range of a semiconductor photo-detector could be enhanced by sensitising the device to sub-band-gap light and by enhancing the photo response for high-energy photons in the same way as described above for a photovoltaic device. Brief Description of the Drawings

Embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings in which:

Figure 1 illustrates a solar cell construction comprising a photovoltaic device and luminescence converter according to one or more embodiments of the present invention wherein the luminescence converter may be an up-converter or a down- converter;

Figure 2 is a schematic energy level diagram showing energy levels in the luminescence-converter of Figure 1. The arrows on the left side (24, 25, 28) indicate the processes involved in an up-conversion process, the arrows on the right side (26, 27, 29) the processes involved in the down-conversion of a high-energy photon. The same three-band material can thus be used either as an up-converter or as a down- converter.

Figure 3 is a schematic energy level diagram showing energy levels in the up- converter of Figure 1 with relaxation of electrons in an intermediate band; Figure 4 is a schematic energy level diagram showing energy levels in the up- converter of Figure 1 with relaxation of electrons in the conduction band; Figure 5 illustrates a solar cell construction with a luminescence converter located on the front surface of a photovoltaic device. According to one embodiment of the present invention the luminescence converter acts as a down-converter;

Figure 6 graphically represents the upper limits for the solar energy conversion efficiency of the up-converting system of Figure 5 as a function of the band-gap for the minimum emission case (open squares) and for. a system with no restriction of the emission angle for concentration factors corresponding to one, 100 and 46200 suns, respectively (circles);

Figure 7 graphically represents the efficiency of the system involving a luminescence converter that acts as a down-converter. Non-concentrated radiation from a 6000K sun and a refractive index n=3.6 for the solar cell and for the luminescence converter was assumed in all calculations, the graphed curves respectively showing:

(Solid line) Shockley-Queisser limit for a conventional solar cell; (Open squares) Efficiency of the system with the converter located on the front surface;

(Solid squares) Efficiency of the system with a converter with one impurity located on the rear surface;

(Solid circles) Efficiency according to a modified Shockley-Queisser model for an infinite number of impurities;

Figure 8 graphically represents the efficiency of the system of Figure 5 as a function of the refractive index n with the converter on the front surface and with band- gap of the solar cell material of 1.1 eV; and

Figure 9 illustrates a solar cell construction in which a down-converter is stacked on the front surface of a photovoltaic device in a similar arrangement to that of Figure 5 and an up-converter is located on the rear surface of the photovoltaic device similar to the arrangement of Figure 1. Detailed Description of Embodiments of the Invention Luminescence-converter acting as an up-converter Referring now to Figures 1 and 2, at least a first embodiment of the present invention provides a solar cell construction 12 comprising a photovoltaic device 13 and a luminescence converter 15, wherein the luminescence converter is an up-converter and the photovoltaic device 13 and the up-converter 15 are electronically isolated from each other by an insulator 14. In cases where the material of the photovoltaic device 13 and of the up-converter 15 do not couple electronically to each other, the insulator 14 can be omitted. Light 11 incident on the solar cell construction includes sub-band-gap light which is transmitted by the photovoltaic device 13 This sub-band-gap light is partially up-converted into high-energy photons by the up-converter 15. The up- converted high-energy photons are subsequently absorbed in the photovoltaic cell 13. Because the converter 15 will emit light omnidirectionally, a reflector 16 is located behind the up-converter 15 to reflect light emitted from the converter 15 back towards the photovoltaic device 13 where the reflected high-energy photons are absorbed in the photovoltaic device 13

Referring to Figure 2, within the luminance converter 15, a photon of low energy will excite electrons from states of low energy 23 to an intermediate state 22 in a first intermediate transition 24. A second low energy photon will then in a second intermediate transition 25 excite the electron from the intermediate state 22 to a higher energy state 21, from which the electron will recombine via a band-to-band transition 28 back to its initial state 23 emitting a high-energy photon.

It might seem that this process could, in principle, proceed with 100% yield, in terms of photon usage. However, each absorption process has its opposite emission process that makes it impossible for electrons to be excited to and from the intermediate state without some recombining back to their originating states. During these recombination processes involving the intermediate state 22 low-energy photons are emitted, which cannot be utilised by the photovoltaic cell 13. The fraction of these photons that is emitted into angles lying inside the escape cone of the photovoltaic cell 13 is lost to the conversion process.

It is found that high up-conversion yield with the standard process involving only the two intermediate transitions 24, 25 and the band-to-band recombination 28 is only possible, even in principle, when the device is operating at high light intensities such as with concentrated sunlight However, it can be shown (see the following theoretical description of the system) that high efficiency is also possible at lower light intensities (e g non-concentrated sunlight) if a relaxation process that wastes some photon-energy by energy relaxation is included in the up-conversion process.

For example, the relaxation could occur when the electron is excited to the intermediate state 122, as in Figure 3. Such relaxation is well known in the area of semiconductor physics where intermediate levels might be formed by impurities or defects within the semiconductor, K. W. Boer Survey of semiconductor physics, Van Nostrand Reinhold, New York, 1990 Upon excitation of an electron to a defect site, the presence of the electron stimulates a physical relaxation in the arrangement of atoms in the vicinity, lowering the energy of the defect level. Similar relaxation processes occur in other optical materials such as dyes During the relaxation process, energy is lost to the neighbouring atoms in the form of phonons

To improve the up-conversion process, the relaxation of the intermediate level should decrease the optical coupling to the ground state 23 but increase this coupling to the final excited state 21 Best results are obtained if the relaxation causes the relevant optical excitation processes to change from being "allowed" to "forbidden" under quantum mechanical selection rules, and vice versa

The relaxation could also occur in the other electron states For example, the relaxation could occur in the final excited state 121 as in Figure 4 In this case, the relaxation should decrease the optical coupling between the intermediate state 22 and this excited state 121 Similarly the relaxation could occur when the electron returns to the ground state 23 Significant improvements of the efficiency of a photovoltaic device are expected for non-concentrated light, when an improved up-converter is used The up-converter 15 in the first embodiment of the invention is therefore preferably an improved up-converter as described above

In a preferred form of the first embodiment of the invention, the photovoltaic device and the up-converter are not spatially separated from each other Instead of a separate up-conversion layer 15, which is located on the rear surface of the photovoltaic device 13 the up-converter is located directly inside the photovoltaic device itself This form of the embodiment is very similar to an impurity photovoltaic device with the major difference that the electrons in the up-conversion material, which is implanted into the photovoltaic device, do not couple strongly to the electronic states in the host material

Several rare earth metal- and transition metal-systems are well known for their preferable optical properties to act as up-converters, D Gamelin, H U G Gudel, Top Curr Chem , 214, 2001 The advantage of this class of materials is, that a completely filled outer electronic shell shields the electrons, which are involved in optical transitions The optical properties of these material systems materials are therefore only marginally affected, when implanted in different host-materials As high up-conversion efficiencies have been demonstrated with rare earth and transition metal systems in other host materials, these systems could also implanted into a photovoltaic device The second preferred form of the first embodiment of the invention therefore consists of a rare earth- or transition metal system or any other material system that can act as an up-converter, that is located directly inside a photovoltaic device The up- conversion system should be located close to the rear surface of the photovoltaic device in order to avoid the absorption of high-energy photons by the up-converter Main advantages of the second preferred form of the first embodiment of the invention are that it can be realised with any given photovoltaic device and that contrary to the first form of the first embodiment of the invention shown in Figure 1, where the up-converter is a separate layer 15 located on the rear surface of the photovoltaic device 13, no bifacial photovoltaic device is required. Luminescence-converter acting as a down-converter

Referring now to Figures 2 and 5, a second embodiment of the present invention provides a solar cell construction comprising a photovoltaic device and a luminescence converter 18 located on the front surface of the photovoltaic device 13, wherein the luminescence converter 15 acts as a down-converter and the photovoltaic device 13 and the down-converter 18 are electronically isolated from each other either by an insulating layer 17 or by virtue of the material characteristics of the converter and photovoltaic device.. Light 1 1 incident on the solar cell construction 212 includes photons of an energy greater than twice the band gap of the photovoltaic device 13. These photons are down-converted into two lower energy photons by the converter 18. The down-converted lower energy photons are subsequently absorbed in the photovoltaic device 13. A reflector 16 is located behind the solar cell 13 to reflect light that passes through the photovoltaic device providing a further opportunity for absorption. In another configuration of the second embodiment of the invention the down- converter is located on the rear surface of the photovoltaic cell as shown in Figure 1. The advantage over the Figure 5 configuration of the embodiment is that higher limiting efficiencies are expected. The emission by the down-converter 18 via the front surface represents a loss mechanism in the Figure 5 configuration of the second embodiment of the invention, which is not present if the down-converter is located on the rear surface of the solar cell. The photovoltaic cell 13 must be transparent to the high energy-part of the incident light in the Figure 1 configuration of the second embodiment of the invention.

A third embodiment of the invention is shown in Figure 9. The device shown is a device incorporating the features of the devices of Figures 1 (up-converter)and Figure 5 (down-converter). The device of Figure 9 provides a solar cell construction 112 comprising a photovoltaic device 13 and luminance converter 18, again located in front of the photovoltaic device, as in Figure 5, wherein the luminance converter 18 is a down-converter and the photovoltaic device 13 and the down-converter 18 are electronically isolated from each other by an insulator 17. A second luminance converter 15 is provided behind the photovoltaic device 13, wherein the luminance converter 15 is an up-converter and the photovoltaic device 13 and the up-converter 15 are electronically isolated from each other by an insulator 14. Light 1 1 incident on the solar cell construction 1 12 includes sub-band-gap photons which are transmitted by the first luminance converter 18, the photovoltaic device 13 and are partially up-converted into high-energy photons by the luminance converter 15. These up-converted high- energy photons are available for subsequently absorption by the photovoltaic device 13. Because the converter 15 will emit light omnidirectionally, a reflector 16 is located behind the up-converter 15 to reflect light emitted from the converter 15 back towards the photovoltaic device 13 where the reflected up-converted high-energy photons are also available for subsequent absorbtion by the photovoltaic device 13. Light 11 incident on the solar cell construction 1 12 also includes photons of energy greater than the band gap of the photovoltaic device 13 and these may be down-converted into two lower energy photons by the converter 18. These down-converted lower energy photons are also potentially available for subsequent absorbtion by the photovoltaic device 13. Finally, incident light 11 of lower energy than those absorbed and converted by the luminance converter 18, but of higher energy than the band gap of the photovoltaic device, pass through the luminance converter 18 and are available for direct conversion to photo-current by the photovoltaic device 13

The following analysis explores the possibility to improve the efficiency of photovoltaic devices in combination with an up-converter in some detail for the ideal case where all parasitic loses have been eliminated. The benefits of energy relaxation for the efficiency of the photovoltaic device are clearly demonstrated. Theoretical description-calculation of the limiting efficiency

The theory for the calculation of the limiting efficiency of the proposed system of a solar cell in combination with a luminescence converter will now be described.

The efficiency of a photovoltaic cell is calculated from its current-voltage characteristics (IV-curve). According to a model introduced by Shockley and Queisser, W. Shockley and H. J. Queisser, J.Appl.Phys., 32, 510 (1961), the IV-curve of a photovoltaic cell is given as the difference between the absorbed and the emitted photon current. The emitted photon current as a function of voltage can be calculated, under the idealising assumptions of no free carrier absorption and infinite carrier mobilities, by a generalisation of KirchhofFs law, T. Trupke, E. Daub and P. Wϋrfel, Sol.Energy Mat.Sol.Cells, 53, 103 (1998).

The absorbed photon current density is given by the integral over the absorptance multiplied by the incident photon current. In limiting efficiency calculations for conventional photovoltaic devices the incident photon current consists of two contributions, one which describes the direct illumination of the device by the sun and a second, hemispherical term, which takes into account that the photovoltaic device receives thermal 300K radiation from the surroundings. In limiting efficiency calculations for a photovoltaic system involving a three-band luminescence converter, one must take into account a third term that describes the luminescence emitted by the luminescence converter. Photon selectivity

The luminescence converter is described as a three-band system. Three types of transitions are possible, namely band-to-band transitions between the valence band and the conduction-band and two types of intermediate transitions between the valence band and the intermediate state and between the intermediate state and the conduction band, respectively. In limiting efficiency calculations for a photovoltaic system involving such a multiple band-system, photon selectivity must be assumed. A lower and an upper threshold energy, Ei and E_m, respectively, must be assigned to each type of transition and the absorptance is assumed to be one for photon energies within these energy limits, and zero for energies outside these energy-limits.

Photon selectivity implies that the energy intervals between the lower- and the upper threshold energy do not overlap for different types of transitions. In that case the luminescence from a three-band system can easily be described as each type of transition can be treated as a separate two-band system and may then be described individually by the generalised KirchhofFs law. Refractive index

As further discussed below, the refractive index n of the photovoltaic cell and of the luminescence converter is a very crucial quantity in some embodiments of the invention. The luminescence emitted by a device via one of its surfaces into the air is the fraction of the spontaneously emitted photons, which are emitted into the direction of that surface and which lie inside the escape cone of that surface. The fraction of photons lying inside the escape cone decreases quadratically with increasing refractive index, which corresponds to an increasing fraction of spontaneously emitted photons, which are totally reflected at the surface. In the generalised KirchhofFs this is accounted for by the fact that the emitted photon current is linear to an effective solid angle, which is Ω=π/n² for a device of refractive index n emitting into the air (n=l).

Contrary, no total internal reflection takes place at an interface between two different materials with equal refractive index. The photon current density emitted by one material with refractive index n into another material with the same refractive index is therefore described by an effective solid angle Ω=π. Theoretical description of different embodiments of the invention

The above-described theoretical calculation of the limiting efficiency can be carried out in very similar ways for different embodiments of the invention due to the fact that the same three-band-system can act, in principle as an up-converter or as a down-converter, depending on the geometry of the system and on the illumination conditions. In all embodiments of the invention a solar cell is working at its maximum power point and this solar cell receives luminescence from the luminescence converter.

A main difference in the description of the different embodiments of the invention are the effective solid angles Ω into which the photovoltaic device and the luminescence converter, respectively emit light, depending on whether the photovoltaic cell or the luminescence converter is located on the front surface. The other main parameters, which must be varied and optimised in the description of different embodiments of the system, are the lower threshold energy and the upper threshold energy for all types of transitions. Different cell geometries and illumination conditions:

Maximum conversion efficiency of sunlight is realised in a system in which the effective solid angle into which luminescent radiation from a solar energy converter can be emitted is equal to the solid angle from which solar radiation is received, i.e.

Ω_s =Ω (7)

This condition can be met in different ways. One possibility is the restriction of the effective external solid angle into which luminescence can be emitted (P.Wurfel, Physik der Solarzellen, 2^nd edition, Spektrum AkadNerl., 2000 and A.Marti, J.L.Balenzategui, R.F.Reyna; J.Appl.Phys.; 82 (8); 1997; 4067). In this situation, which will be denoted the minimum emission case in the following, luminescent emission is possible only into the small solid angle made up by the solar disc.

Another option is to focus the sunlight onto the surface of the solar cell with an infinitely extended lens. This is the maximum concentration case with an incident energy current density that is 46200 times larger than with non-concentrated light.^'

In the limiting efficiency calculations described below, the sun was modelled as a T=6000K blackbody radiator and the photovoltaic cell and the luminescence converter were assumed to be at T=300K. Results

Up-converter located on the rear surface of a photovoltaic device.

The limiting conversion efficiency is calculated for different illumination conditions for the first embodiment of the invention, which involves an up-converter located on the rear surface of a photovoltaic device. In these calculations one assumes equal band-gaps of the photovoltaic cell 13 and of the up-converter 15.

Figure 6 shows the limiting efficiency as a function of the band-gap of the photovoltaic cell 13 and of the up-converter 15. For each value of the band-gap the position of the intermediate level has been optimised with respect to the efficiency of the system.

The highest efficiency is obtained in the minimum emission case (open squares in Figure 6). For a band-gap of 1.955 eV and an energy of the intermediate level at Eι=0.713eV the maximum efficiency achievable with equal band-gaps of the solar cell and of the up-converter is 63.17%. The maximum-efficiency for the maximum concentration case (circles, upper curve) is slightly lower. For E_g=1.86 eV and an intermediate level at Eι=0.667 eV a maximum efficiency of 61.40 % is calculated.

The maximum efficiency for other concentration-factors corresponding to one and 100 suns are also plotted in Figure 6 (circles). These curves indicate that the maximum achievable efficiency of the system decreases strongly with decreasing illumination intensity. For comparison the Shockley-Queisser limits (W. Shockley) of conventional single junction solar cells are also plotted in Figure 6 for the same concentration factors (solid lines). While the efficiencies calculated for the first embodiment of the invention are slightly above the Shockley-Queisser efficiency even for non-concentrated light, Figure 6 suggests that a significant improvement of the efficiency is expected only for concentrated light. This shows that with an up- converter involving exclusively the two intermediate transitions (24, 25 in Figure 2) and the band-to-band transition 28 the up-conversion efficiency, i.e. the fraction of incident low-energy photons, which are up-converted into high-energy photons, is very low.

Effect of relaxation on the maximum efficiency - improved up-conversion

A significant improvement of the up-conversion efficiency and consequently of the efficiency of the photovoltaic system, is achieved for non-concentrated light if relaxation processes are allowed whereby electrons (or holes) in excited states can relax to slightly lower lying states, which couple only weakly to the level from which the electron (hole) was excited. Such a relaxation is present e.g. in dye molecules where a radiative transition between two singlet states is often followed by a very fast relaxation process into a lower lying triplet state (intersystem crossing). Various defect centres in semiconductors also relax after electron capture by a change in the local atomic arrangement (K.W.Bόer; Survey of semiconductor physics; Van Nostrand Reinhold, New York; 1990; p. 495). Corresponding energy level diagrams are shown in Figure 3 in which energy levels in the up-converter 15 are illustrated with relaxation of electrons in the conduction band 121 and in which radiative electronic transitions involving the intermediate level 22 are only allowed between the intermediate level and the dark shaded energy regions of the bands 23, 121 whereas band-to-band transitions also involve the relaxed states in the lightly shaded energy region of the conduction band 121. In Figure 4 the relaxation is illustrated as taking place between different intermediate states inside an intermediate band 122.

In the limit where the coupling after the relaxation changes to very weak one can simply implement such a relaxation and change in the corresponding selection rules into the model by allowing the sum of the energy differences between the valence-band edge and the intermediate state and between the intermediate state and the conduction- band edge to be larger than the lower threshold energy for the band-to-band transitions. Thus, referring to Figure 3 and 4,

E_l + E₂ = E_g + E_relax where E_reiax is the relaxation energy.

The inclusion of this relaxation into the calculations results in a significant improvement of the efficiency of the system. A maximum value of 47.6% is calculated for Eg=2 eV and for a relaxation energy of 0.33 eV triangles in Figure 6). Including a relaxation into the calculations for the maximum concentration case yields a limiting efficiency of 62.18 % for Eg=1.9eV, Eι=0.7206 eV, E₂=1.2294 eV and a relaxation energy of 50 meV. In the minimum emission case a relaxation of the charge carriers does not lead to a limiting efficiency higher than the 63.11% reported above. Relevance of the refractive index

The integrated photon current emitted by the up-converter via the two intermediate transitions and via band-to-band transitions was calculated. For the different illumination conditions described above, it is found that the photon current spontaneously emitted by the up-converter via the intermediate transitions is larger than the spontaneous emission of up-converted photons via band-to-band transitions, which seems to contradict the high efficiencies which are calculated for these cases. The origin of this discrepancy is the fact that in the first embodiment of the invention the up-converted photons are spontaneously emitted by the up-converter 15 into the photovoltaic device 13, i.e. into a material with equal refractive index, which according to the above considerations is described by Ω=π. The low-energy-photons emitted by the up-converter 15 are transmitted by the photovoltaic cell 13. Due to the large refractive index it is assumed in calculations that (n=3.6, typical for silicon) the effective solid angle for the low-energy-photons, which are able to leave the system via the front surface is only Ω=π/n². The major fraction of the low energy photons is totally reflected at the front surface of the photovoltaic cell 13 and can then be recycled inside the up-converter 15 to get a further chance to be up-converted. In the maximum concentration case the recycling factor for sub-band-gap photons, i.e. the ratio of the internal effective solid angle to the solid angle into which emission from the front surface takes place is only n², whereas it is 46200 times n² in the minimum emission case. The latter figure for the recycling factor is much larger than necessary for high efficiencies. On the other hand a change of the refractive index from n=l o n=3.6 yields a very large boost of the efficiency in the maximum concentration case. A large and similar refractive index for the luminescence converter and for the photovoltaic device is therefore essential in the maximum concentration case. Photon selectivity Photon selectivity can be achieved in a three-band system as shown in Figure 2 by limiting the widths of the conduction- and of the valence band as schematically shown in Fig.2. The energy-difference between the lower valence-band-edge and the upper conduction-band edge is ideally E_g+ E₂.

The same energy scheme has to be used in an IPV-solar cell in order to ascertain photon selectivity. However, in an EPV-system this energy-scheme sets a limitation to the maximum conversion-efficiency, as solar photons with energies larger than E_g+ E cannot be absorbed. Instead of the maximum efficiency of 63.2%, which has been reported as the limiting efficiency of the IPV solar cell for maximal concentration, A. Luque and A. Marti, Phys. Rev. Lett., 78 (26), 5014 (1997), the upper limit of the _. conversion efficiency is 58.9% under these more realistic constraints; A. Brown, M. A. Green and R. P. Corkish; Int. Workshop on Nanostructures in Photovoltaics; Dresden; 2001, to be published in Physica E.

In the first embodiment of the invention the incident high-energy photons are absorbed by the solar cell 13, in which the widths of the bands are not limited. Restricting the widths of the bands in the up-converter 15 as shown in Figure 2 therefore doesn't represent a loss. Down-converter located on the rear surface of a solar cell

Referring to Figure 1, in one preferred form of the second embodiment of the invention a down-converter is located behind a photovoltaic device that is transparent to photons with energies larger than twice the band-gap-energy. In this embodiment of the invention the emission of high-energy photons by band-to-band transitions inside the converter is not wanted. The calculations show that for a non-concentrated 6000K blackbody spectrum the integral photon current emitted via the two intermediate transitions is more than 5000 times larger than the integral photon current emitted via band-to-band transitions. As no recycling of the few emitted high-energy photons is necessary, the refractive index of the down-converter may be lower than the refractive index of the photovoltaic cell.

The limiting efficiency for the second embodiment of the invention is shown in Fig.7, solid squares. A maximum efficiency of 39.55 % is found for a band-gap-energy of 1.1 eV, which makes this approach particularly interesting for silicon solar cells. Down-converter on the front surface of the system:

Referring now to Figure 5, another preferred form of the second embodiment of the invention consists of a geometry with the luminescence converter located on the front surface of the system. Incident light with energies bω > Ei is completely absorbed by the down-converter and the solar cell only receives the luminescence emitted by the down-converter.

The efficiency of that system is shown as open squares in Figure 7. The curve is significantly below the efficiency curve calculated for a system with the down- converter on the rear surface (solid squares). These deviations are due to the emission of photons by the down-converter via the front surface. The light which is emitted by the luminescence converter via the front surface is lost if the converter is located on the front surface of the solar cell. This additional loss mechanism is also responsible for the fact that the efficiency of the down-conversion system with a down-converter on the front surface is below the Shockley-Queisser limit for a conventional solar cell at large band-gap-energies (see Figure 7). If the down-converter is located at the front surface of the solar cell one might expect that only half of the luminescence emitted by the down-converter contributes to the photocurrent of the solar cell. This would in fact be the case for a material with a refractive index n=l which is emitting hemispherically. However, solar cell materials like silicon or GaAs have a large refractive index of typically n«3.6. As mentioned above, the effective solid angle into which luminescence is emitted by the down- converter via the front surface into the air is Ω=π/n², while the emission from the down-converter in the direction of the solar cell is going into Ωj_nt=π, hence a 13 times larger solid angle for n=3.6.

Only the fraction of photons, which is emitted into the direction of the solar cell, contributes to the photocurrent of the cell. Therefore a large refractive index of the solar cell and of the up-converter is essential for a high conversion efficiency of the system if the converter is located on the front surface. As an example the efficiency of a solar cell with a down-converter on the front surface is shown as a function of the refractive index n (equal for the solar cell and the luminescence converter) and for a band-gap-energy of 1.1 eV in Figure 8. As expected the efficiency drops below the Shockley-Queisser limit as n approaches one. However, with increasing refractive index n the photon current emitted via the front surface converges against zero and the efficiency therefore converges towards 39.55% the limiting value which is calculated for the embodiment of the invention, where the down-converter is located on the rear surface of the solar cell. Relaxing the assumption A(bω)=l:

Increasing the refractive index above n=3.6 is certainly not a practical way to bring the efficiency of a real system closer to the limiting efficiencies possible with the luminescence converter located on the rear surface of the solar cell (solid squares in Figure 7). However, a further improvement can be achieved if one relaxes the assumption A(bω)=l for the two intermediate transitions, while this condition is maintained for the band-to-band transitions. One disadvantage of the second form of the second embodiment of the invention with the converter on the front surface as described so far is that not only the high-energy photons but also the lower energy photons with El< bω< E1+E2, which are not suitable for the down-conversion process are completely absorbed by the down-converter. The subsequent re-emission of photons results in a randomisation of the direction of the photons that would otherwise completely be absorbed by the solar cell. The fraction of photons, which is emitted into the escape cone of the front surface of the down-converter, is lost. By reducing the absorptance for the intermediate transitions to e.g. A(bω)=0.01 for El< bω< E1+E2 the fraction of incident photons in this energy interval, which is absorbed by the down- converter is largely reduced. According to KirchhofFs law the photon current emitted by the down-converter is linear in the absorptance. The reduction of A(bω) for the intermediate transitions thus also reduces the ratio of the photon current emitted by the intermediate transitions over the photon current emitted by band-to-band transitions. However, as mentioned above, this ratio is approximately 5000 if all transitions have total absorption. Therefore, the major fraction of photons is still emitted by the intermediate transitions even if the down-converter is almost transparent in the corresponding energy interval.

The advantage of this modification of the system is that the incident sunlight with photon energies El< bω< E1+E2, which cannot be down-converted, is almost completely transmitted by the down-converter and absorbed in the solar cell and thus does not suffer from the additional loss mechanism due to the luminescence converter. An upper limit for the efficiency of 38.6% is calculated for a band-gap of 1.1 eV for the first form of the second embodiment of the invention in this case. Photon selectivity: In the theoretical calculations described above photon selectivity has been assumed. In a three-band system this can be achieved e.g. by limiting the width of the bands. However this restriction of the bands sets an upper threshold energy for the absorptance of the converter and thus reduces the number of photons that can be down- converted, which would result in a lower efficiency of the system. A further advantage of reducing the absorptance of the intermediate transitions is that a restriction of the bands of the luminescence converter is no longer necessary to achieve photon selectivity. The low absorptance guarantees that only a minor fraction of high-energy photons is absorbed via one of the intermediate transitions.

Compared to the impurity photovoltaic effect or the use of impact ionisation the presented embodiments of the invention have advantage that the efficiency of a photovoltaic cell is improved by adding a component to an existing solar cell, which can be optimised independently from the solar cell material itself. No additional constraints are imposed on the solar cell material like e.g. the exclusion of electron- phonon coupling as in the case of impact ionisation or like the need to find a suitable impurity for a particular solar cell material in an IPV solar cell.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

1 A photovoltaic device comprising a photovoltaic cell, a luminescence converter and a reflector

2 The device of claim 1, wherein the luminescence converter is an up-converter, which transforms sub-band-gap photons transmitted by the photovoltaic cell to higher energy-photons, which are capable of subsequently being absorbed by the photovoltaic cell, the up-converter being located between the photovoltaic cell and the reflector, which thereby functions as a rear reflector, and the up-converter is optically coupled to the photovoltaic cell 3 The device of claim 2, wherein the up-converter and the photovoltaic cell are substantially electronically isolated from each other

4 The device of claim 2, wherein the up-converter and the photovoltaic cell are only weakly coupled electronically to each other

5 The device of claim 2, 3 or 4, wherein the up-converter and the photovoltaic cell are each fabricated from materials which do not support strong electrical coupling between the up-converter and the photovoltaic cell when the up-converter and the photovoltaic cell are in direct mechanical contact

6 The device of claim 2, 3, 4 or 5, wherein the up-converter and the photovoltaic cell all have substantially similar refractive indices 7 The device of claim 2, 3 or 4, wherein the up-converter and the photovoltaic cell are separated from each other by a transparent insulating layer

8 The device of claim 7, wherein the up-converter, the photovoltaic cell and the insulating layer all have substantially similar refractive indices

9. The device as claimed in any one of claims 2 to 8, wherein the up-converter consists of a system in which three bands are involved in optical transitions, and where each band consists of allowed energy states of a single energy or spread over a range of energies.

10 The device as claimed in any one of claims 2 to 9, wherein the up-converter employs an up-conversion processes known as ground state absorption/ excited state absorption (GSA/ESA), which involves the generation of an excited state by a two-step process involving a real, lower-lying, metastable state or states

11 The device as claimed in claim 10, wherein the up-converter includes at least one excited energy state to which electrons can be excited from an initial lower energy state by absorption of an incident sub-band-gap energy photon and from which an electron will relax into a rest state which is a lower energy state than the excited state but higher than the initial state with consequent release of a small amount of energy and where the rest state is only weakly linked to the initial state from which the electron was originally excited

12 The device as claimed in claim 11, wherein the rest state is associated with the upper band 13 The device as claimed in claim 11, wherein the rest state is associated with an intermediate level band

14 The device as claimed in claim 11, wherein both the upper band and an intermediate band have an associated rest state

15 The device of as claimed in any one of claims 2 to 14, wherein the up-converter incorporates a semiconductor or any other material with a band-type electronic density of states and with an intermediate level or intermediate band within the band gap where all transitions are radiatively efficient, to act as a broad band up-converter

16 The device as claimed in claim 1, wherein the luminescence converter is a down-converter which absorbs photons at an energy at or greater than twice the band gap of the photovoltaic cell and emits two or more photons at energies at or above the band-gap energy of the photovoltaic cell, whereby the lower-energy photons are each capable of being absorbed by the photovoltaic cell to contribute to the photocurrent of the photovoltaic cell

17 The device as claimed in claim 16, wherein the down-converter is located in front of a photovoltaic cell and the reflector is located behind the photovoltaic cell, whereby illumination must pass through the down-converter to reach the photovoltaic cell

18 The device as claimed in claim 16, wherein the down-converter absorbs photons with energies equal to or larger than twice the band-gap energy E_g of the photovoltaic device and is substantially transparent to photons with energies larger than E_g but smaller than 2E_g

19 The device as claimed in claim 16 or 17, wherein the down-converter and the photovoltaic cell are optically coupled to each other but substantially electrically isolated from each other 20 The device as claimed in claim 19, wherein the down-converter and the photovoltaic cell are separated by an interspaced transparent insulating layer

21 The device as claimed in claim 20 wherein the down-converter, the photovoltaic cell and the insulating layer all have substantially similar refractive indices

22 The device as claimed in claim 16, 17, 18, or 19, wherein the down-converter and the photovoltaic cell are directly mechanically attached to one another and are not electrically coupled 23 The device as claimed in claim 16, wherein the down-converter is located on the rear surface of the photovoltaic cell

24 The device as claimed in claim 23, wherein the photovoltaic cell has a band-gap energy of E_g=l 1 eV 25 The device as claimed in claim 23 or 24, wherein for the photovoltaic device is a silicon solar cell or a silicon photo detector

26 The device as claimed in any one of claims 16 to 25, wherein the down- converter uses a material with a separation between the upper and lower band of states equal to at least 2 E_g, which contains states or bands of intermediate energies 27 The device as claimed in any one of claims 16 to 26, wherein the material of the down converter employs absorption of high-energy photons by a band-to-band transition which leads to the generation of electron-hole-pairs inside the luminescence converter and the radiative recombination of these electron-hole-pairs via the impurity level is accompanied by the emission of two lower energy photons 28 The device as claimed in any one of claims 16 to 27, wherein the down converter uses rare earth compounds which enable conversion of ultraviolet light into visible light

29 The device as claimed in claim 1, wherein an up-converter is located on a rear surface of the photovoltaic cell between the reflector and the photovoltaic cell, a down- converter is located on the front surface of the photovoltaic cell, whereby the up- converter and the down-converter are combined to reduce both, the transmission losses and the thermalisation losses of the device

30 The device of claim 29, wherein the up-converter and the photovoltaic cell are substantially electronically isolated from each other 31 The device of claim 29, wherein the up-converter and the photovoltaic cell are only weakly coupled electronically to each other

32 The device of claim 29, 30 or 31, wherein the up-converter and the photovoltaic cell are each fabricated from materials which do not support strong electrical coupling between the up-converter and the photovoltaic cell when the up-converter and the photovoltaic cell are in direct mechanical contact

33 The device of claim 29, 30, 31 or 32 wherein the up-converter and the photovoltaic cell all have substantially similar refractive indices

34 The device of claim 29, 30 or 31, wherein the up-converter and the photovoltaic cell are separated from each other by a transparent insulating layer 35 The device of claim 34, wherein the up-converter, the photovoltaic cell and the insulating layer all have substantially similar refractive indices

36. The device as claimed in any one of claims 29 to 35, wherein the up-converter consists of a system in which three bands are involved in optical transitions, and where each band consists of allowed energy states of a single energy or spread over a range of energies. 37. The device as claimed in any one of claims 29 to 36, wherein the up-converter employs an up-conversion processes known as ground state absorption/ excited state absorption (GSA/ESA), which involves the generation of an excited state by a two-step process involving a real, lower-lying, metastable state or states.

38. The device as claimed in claim 37, wherein the up-converter includes at least one excited energy state to which electrons can be excited from an initial lower energy state by absorption of an incident sub-band-gap energy photon and from which an electron will relax into a rest state which is a lower energy state than the excited state but higher than the initial state with consequent release of a small amount of energy and where the rest state is only weakly linked to the initial state from which the electron was originally excited.

39. The device as claimed in claim 38, wherein the rest state is associated with the upper band.

40. The device as claimed in claim 38, wherein the rest state is associated with an intermediate level band. 41. The device as claimed in claim 38, wherein both the upper band and an intermediate band have an associated rest state.

42. The device of as claimed in any one of claims 29 to 41, wherein the up- converter incorporates a semiconductor or any other material with a band-type electronic density of states and with an intermediate level or intermediate band within the band gap where all transitions are radiatively efficient, to act as a broad band up- converter.

43. The device as claimed in any one of claims 29 to 42, wherein the down- converter is located in front of a photovoltaic cell and the reflector is located behind the photovoltaic cell, whereby illumination must pass through the down-converter to reach the photovoltaic cell.

44. The device as claimed in any one of claims 29 to 42, wherein the down- converter absorbs photons with energies equal to or larger than twice the band-gap energy E_g of the photovoltaic device and is substantially transparent to photons with energies larger than E_g but smaller than 2E_g. 45 The device as claimed in any one of claims 29 to 44, wherein the down- converter and the photovoltaic cell are optically coupled to each other but substantially electrically isolated from each other

46 The device as claimed in claim 45, wherein the down-converter and the photovoltaic cell are separated by an interspaced transparent insulating layer

47 The device as claimed in claim 46, wherein the down-converter, the photovoltaic cell and the insulating layer all have substantially similar refractive indices

48 The device as claimed in any one of claims 29 to 45, wherein the down- converter and the photovoltaic cell are directly mechanically attached to one another and are not electrically coupled

49 The device as claimed in any one of claims 29 to 48, wherein the down- converter uses a material with a band-gap E_g, which contains states or bands with energies inside the band-gap 50 The device as claimed in any one of claims 29 to 49, wherein the material of the down converter employs absorption of high-energy photons by a band-to-band transition which leads to the generation of electron-hole-pairs inside the luminescence converter and the radiative recombination of these electron-hole-pairs via the impurity level is accompanied by the emission of two lower energy photons 51 The device as claimed in any one of claims 29 to 50, wherein the down converter uses rare earth compounds which enable conversion of ultraviolet light into visible light

52 The device as claimed in any one of the preceding claims when used as a photo- detector